 |
 |
Ylva If you created the images in spm5, they will not be properly read in spm99, i.e., the image data can be read but the header and the transformation matrix will not be read properly. This will cause a misalignment between your image and the glass brain. The images will look OK if you display them using the display button, as it centers the images. I'm not sure why your design would not work in spm5. It looks like repeated measures anova with 1 within and 1 between subject factor. This should be relatively straightforward to setup using the flexible factorial option. ----- Darren On Tue, Oct 28, 2008 at 6:50 PM, Ylva Lilja <[log in to unmask]> wrote: > Dear SPM experts, > > We are having trouble doing a group by condition anova in SPM. We tried > different ways in SPM5 (flexible factorial), without success, and are now > trying in SPM99, Multi-group: condition & covariates. We have two groups and > two conditions. We get the design attached, which looks satisfactory: > > The contrast images (1 -1 -1 1 and -1 1 1 -1) look realistic, when we > eye-ball them in Vinci. > > Our problem is that at the very end where the points of significant > difference in variance are seen outside the glass brains, as in the 2nd > image attached. It seems like there is something wrong with the coordinates, > for some reason. > > > Has anyone had this problem before? We would be very grateful for any help > with this! > > Best regards, > Ylva Lilja > UC Irvine